Lesson 2 · AI Around Us
Choosing the useful clues that help a machine tell examples apart.
When you learn to spot a cricket fan, you don't memorise every detail about them. You focus on the useful clues.
AI does the same. Those useful clues have a name: features.
A feature is a single useful piece of information that describes an example.
To recommend a sport, good features might be:
Pick the clues that actually help the prediction.
Suppose we want to guess if someone enjoys football:
| Clue | Useful? |
|---|---|
| Enjoys outdoor team games | ✅ Very useful |
| Has lots of energy | ✅ Useful |
| Colour of their school bag | ❌ Useless |
| Their roll number | ❌ Useless |
Good features carry a signal. Useless ones are just noise.
To predict if a student will enjoy football, which feature is the most useful?
You're building an AI to recommend a book to a classmate. List 3 useful features it should notice — and 1 useless clue it should ignore.
Watch out for misleading features. If, by accident, all the cricket fans in your data were boys, the AI might wrongly decide "cricket = boys." That's how unfair bias sneaks in — more on that later.
Name one misleading feature for predicting a student's exam result — a clue that looks helpful but could fool the AI — and explain why it's misleading.
This chapter is free — open this lesson in your portal to save your answers and track progress.